Main circuit design and double-closed-loop PI control method based on Boost circuit

Abstract

The invention belongs to the technical field of wireless power transmission, and particularly discloses a main circuit design and double closed-loop PI control method based on a Boost circuit, which comprises the following steps that a wireless charging system receives fixed-frequency alternating current from a power grid, and the fixed-frequency alternating current passes through a rectifying circuit, a direct-current bus, a PWM (Pulse Width Modulation), a primary compensating circuit, a secondary compensating circuit, a rectifying circuit and a DC / DC converter; the voltage is converted into a required voltage value, a Boost circuit is adopted, and a BP-PI controller is used for calculation. The problem that the charging efficiency is low due to the fact that appropriate current cannot be provided in time due to the fact that the output voltage of a Buck circuit is unstable is solved. An existing Buck circuit is replaced by the Boost circuit, superposition of input voltage and inductive voltage can be achieved in the on-off process of the switching tube, the capacitor is charged, and power is supplied to the load, so that the output voltage is higher than the input voltage, and the voltage conversion efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless power transmission, and in particular to a main circuit design based on a Boost circuit and a double closed-loop PI control method. Background Art

[0002] With the widespread use of electric vehicles, users have put forward higher requirements for driving range, efficiency and safety of charging equipment. In this context, wired charging electric vehicles are often limited by charging piles, and charging equipment may have problems such as aging and poor contact, increasing safety risks and no longer meeting the needs of the public.

[0003] The wireless charging technology for electric vehicles uses power rails buried underground to transmit electrical energy in the form of a high-frequency alternating magnetic field to the power pickup mechanism at the receiving end of a vehicle running within a certain range on the ground, thereby powering the on-board energy storage device, allowing electric vehicles to carry a small number of battery packs and extend their range, while making power replenishment safer and more convenient.

[0004] The Buck circuit in the existing wireless charging system lacks adaptive adjustment capabilities in voltage control, has weak resistance to adverse factors, is difficult to accurately control the output voltage and is unstable. In terms of current response, its control method has limitations and insufficient dynamic performance. It cannot respond to load changes as quickly as specific methods, resulting in the inability to provide appropriate current in time when facing rapidly changing charging needs, ultimately resulting in low charging efficiency. Summary of the invention

[0005] The purpose of the present invention is to provide a main circuit design based on a Boost circuit and a dual closed-loop PI control method to solve the problem that the Buck circuit in the existing wireless charging system has an unstable output voltage and a limited control method for current response, which leads to the inability to provide appropriate current in time and ultimately leads to low charging efficiency.

[0006] To achieve the above object, the basic solution provided by the present invention is: a main circuit design based on a Boost circuit and a double closed-loop PI control method, comprising the following steps:

[0007] S1. The receiving end of the wireless charging system receives fixed frequency AC power from the transmitting end of the power grid and converts it into DC power through a rectifier circuit.

[0008] S2, DC power passes through the DC bus and enters the high-frequency inverter, which changes the current direction by controlling the on and off of D1, D2, D3 and D4, and converts DC power into AC power;

[0009] S3, the AC power is then modulated by PWM pulse width modulation to change the duty cycle to change the frequency of the AC power, converting the DC power into high-frequency AC power;

[0010] S4, high-frequency AC passes through the primary compensation circuit, and the LCC resonant circuit selects the signal at a specific frequency, amplifies or transmits it, and attenuates the signals of other frequencies;

[0011] S5, high-frequency alternating current is transmitted to the receiving winding, and the receiving winding then transmits the high-frequency alternating current to the secondary compensation circuit for filtering;

[0012] S6, the alternating current is then converted into direct current through a rectifier circuit;

[0013] S7. The DC power is then converted into the required voltage value by the DC / DC converter. The Boost circuit is used. When the switch tube is turned on, the input voltage charges the inductor L and the capacitor C discharges the load R. As time goes by, the current on the inductor increases and the voltage across the capacitor C decreases. When the switch tube is cut off, the inductor L discharges. As time goes by, the current on the inductor decreases and the voltage across the capacitor C increases. At this time, the input voltage and the voltage on the inductor L are superimposed to charge the capacitor C and supply power to the load R. By feeding the PWM control pulse to the switch tube, the switch state of the switch tube is continuously changed until the output voltage in the circuit is higher than the input voltage.

[0014] S8. In the continuous conduction switch tube operation mode, there is a negative feedback control loop for voltage and current. The error value is obtained by comparing the output voltage with the reference voltage, and the reference input of the current loop is obtained by calculation by the BP-PI controller. By making the input current track the reference input, the current signal and voltage sampling phase on the input side are adjusted.

[0015] S9, establishing a simulation model for PI double closed-loop control based on BP neural network, and performing boost circuit simulation;

[0016] S10. Finally, analyze the simulation results.

[0017] The beneficial effects of the present invention are as follows: by adopting a Boost circuit to replace the existing Buck circuit, the problems of unstable output voltage and limited current response control mode of the Buck circuit are solved; the Boost circuit can realize the superposition of input voltage and inductor voltage during the on and off process of the switch tube, charge the capacitor and supply power to the load, so that the output voltage is higher than the input voltage, thereby improving the voltage conversion efficiency; and a double closed-loop PI control method is introduced, in which there is a negative feedback control loop for voltage and current, an error value is obtained by comparing the output voltage with a reference voltage, and a reference input of the current loop is calculated by a BP-PI controller, so that the input current tracks the reference input, thereby realizing the control of the input side current signal and the voltage sampling phase, and improving the stability of the system.

[0018] Option 2 is the preferred option of the basic option. In S1, the rectifier circuit adopts a bridge rectifier circuit with an ultrafast recovery diode. When the coupling factor between the transmitter and the receiver is high, the signal can be approximated to contain only the fundamental frequency, and the full-wave signal becomes a half-wave after passing through the rectifier bridge.

[0019] Solution three is a preferred solution of the basic solution. In S4, the bilateral LCC resonant circuit is composed of an LCC resonant compensation circuit composed of capacitors and inductors and a magnetically coupled resonant coil, which is convenient for improving the working efficiency of the system.

[0020] Scheme 4 is the preferred basic scheme. In S7, in the continuous current working mode, the stable working point of the boost circuit satisfies the inductive volt-second balance principle:

[0021] V S DT=(U C -V S )(1-D)T

[0022] Where: V S and U C are the input and output voltages of the boost converter, respectively, D is the duty cycle of the switching signal, and T is the period of the switching signal;

[0023] For a boost converter, the input power and output power of the converter are equal under steady-state operating conditions, and the relationship between the two can be expressed as follows:

[0024]

[0025] Where: i s is the input current of the boost converter, i o is the output current of the boost converter and R is the load resistance.

[0026] Combined with the relationship between the input and output voltage of the boost converter:

[0027]

[0028] In the steady-state operation of the boost circuit, when the switch is turned on, the output capacitor C is responsible for supplying power to the load during the charging phase of the energy storage inductor. The relationship between the energy released by the energy storage inductor and the corresponding change in the output voltage of the energy storage inductor during the switch-on period can be expressed as:

[0029]

[0030] Where: ΔU C is the ripple of the output voltage of the boost converter, and U is the inductor voltage.

[0031] Scheme 5 is the preferred basic scheme. In S8, the PI algorithm used has the following formula:

[0032]

[0033] e(t)=U k -I ref

[0034] Where: K P Indicates the ratio, K i They represent the integral coefficient, e(t) is the input error, U k is the output of the BP-PI controller, I ref is the feedback value of the current loop, and u(t) is the output of the current controller.

[0035] The parameter value of PI cannot be negative. The excitation function of the network output layer uses a non-negative Sigmoid function, and its formula is as follows:

[0036]

[0037] Where: g(x) is the hyperbolic tangent function.

[0038] The excitation function of the hidden layer neurons is allowed to be negative, and the Sigmoid function is selected to be positive and negative symmetric. The formula is as follows:

[0039]

[0040] Take the performance indicator function as an example:

[0041]

[0042] Where: E(k) is the performance indicator function (loss function), which represents the error between the network prediction value and the target value, r t is the target value, y t is the output (prediction value) of the network.

[0043] The input layer of the network can be shown as follows:

[0044]

[0045] in: is the first layer input of the neural network, M is the number of input layers, and in the present invention, M is 8. is the bias node (constant term) of the input layer.

[0046] The input and output of the network hidden layer can be shown as follows:

[0047]

[0048]

[0049]

[0050] in: is the input of the i-th node in the hidden layer, is the weight between the input layer and the hidden layer, is the output of the input layer, is the output of the i-th node in the hidden layer, q is the number of nodes in the hidden layer, is the bias node of the hidden layer (constant term).

[0051] The input and output of the network output layer are as follows:

[0052]

[0053]

[0054]

[0055]

[0056]

[0057] in: is the input of the lth node in the output layer, is the weight between the hidden layer and the output layer, (k) is the output of the hidden layer, is the output of the lth node in the output layer, is the output variable of the network, K P is the control parameter, K t is the temperature related parameter, u k For the control signal.

[0058] Based on the above analysis, the learning formula for the network output layer weight is:

[0059]

[0060]

[0061] in: is the weight adjustment between the lth node of the output layer and the lth node of the hidden layer, α is the momentum term, η is the learning rate, is the error signal of the output layer, e(k) is the error of the loss function, and sgn is the sign function used to determine the gradient direction.

[0062] Learning algorithm for hidden layer weight coefficients:

[0063]

[0064]

[0065] in: It is the error signal of the hidden layer, which represents the combination of the output error of the hidden layer nodes and the corresponding weights.

[0066] Solution 6 is the preferred basic solution. In S9, a BP neural network algorithm is added to the outer loop of the simulated voltage, and the inner current loop uses the basic PI module provided by MATLAB. The inner loop controller outputs a periodic duty cycle signal to control the IGBT.

[0067] Scheme 7, which is the preferred basic scheme, in S10, the simulation target voltage is 30V, the cycle is 0.05s, the output voltage is U0, and the efficiency is calculated by the following formula:

[0068]

[0069] Where: output power is P out , the input power is P in , the efficiency is η. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 is a flow chart of the present invention;

[0071] Figure 2 is a topological diagram of the LCC resonant compensation circuit in the present invention;

[0072] Figure 3 It is a schematic diagram of the switch tube of the transmitter inverter in the present invention;

[0073] Figure 4 is an equivalent circuit diagram when the switch tube in the present invention is turned on;

[0074] Figure 5 is an equivalent circuit diagram when the switch tube is disconnected in the present invention;

[0075] Figure 6 It is the BP-PI double closed-loop control block diagram of the present invention;

[0076] Figure 7 It is the BP neural network structure in the present invention;

[0077] Figure 8 It is the Boost circuit simulation model in the present invention;

[0078] Fig. 9 It is the BP-PI simulation model in the present invention;

[0079] Fig.10 It is the BP-PI module in the present invention;

[0080] Fig.11It is a simulation model of the Boost main circuit of the main circuit in the present invention;

[0081] Fig.12 It is the PI single closed-loop simulation model in the present invention;

[0082] Fig.13 The present invention is a voltage diagram of a Boost circuit;

[0083] Fig.14 The present invention is a Buck step-down circuit voltage diagram;

[0084] Fig.15 The present invention is a diagram of the simulation result of the charging efficiency of the Boost circuit;

[0085] Fig.16 The present invention is a diagram of the simulation results of the charging efficiency of the Buck step-down circuit. DETAILED DESCRIPTION

[0086] The present invention is further described in detail below through specific embodiments:

[0087] like Figures 1 to 16 As shown: A main circuit design based on a Boost circuit and a double closed-loop PI control method, comprising the following steps:

[0088] S1. The receiving end of the wireless charging system receives fixed frequency AC power from the transmitting end of the power grid, and passes through the rectifier circuit. The rectifier circuit adopts a bridge rectifier circuit with ultra-fast recovery diodes. When the coupling factor between the transmitter and the receiver is high, the signal can be approximated to contain only the base frequency. The full-wave signal becomes a half-wave after the rectifier bridge and is converted into DC power.

[0089] S2, DC power passes through the DC bus and enters the high-frequency inverter, which changes the current direction by controlling the on and off of S1, S2, S3 and S4, and converts DC power into AC power;

[0090] S3, the AC power is then modulated by PWM pulse width modulation to change the duty cycle to change the frequency of the AC power, converting the DC power into high-frequency AC power;

[0091] S4, high-frequency alternating current passes through the primary compensation circuit, and the LCC resonant circuit selects the signal at a specific frequency, amplifies or transmits it, and attenuates the signals of other frequencies. The bilateral LCC resonant circuit is composed of an LCC resonant compensation circuit composed of capacitors Lf1 / Lf2 and inductors L1 / L2 and a magnetically coupled resonant coil; Figure 1 , Cf1 / Cf2 are primary / secondary parallel compensation capacitors, C1 / C2 are primary / secondary series compensation capacitors;

[0092] S5, high-frequency alternating current is transmitted to the receiving winding, and the receiving winding then transmits the high-frequency alternating current to the secondary compensation circuit for filtering;

[0093] S6, the alternating current is then converted into direct current through a rectifier circuit;

[0094] S7. The DC power is then converted into the required voltage value by the DC / DC converter. The Boost circuit is used. When the switch tube is turned on, the input voltage charges the inductor L and the capacitor C discharges the load R. As time goes by, the current on the inductor increases and the voltage across the capacitor C decreases. When the switch tube is cut off, the inductor L discharges. As time goes by, the current on the inductor decreases and the voltage across the capacitor C increases. At this time, the input voltage and the voltage on the inductor L are superimposed to charge the capacitor C and supply power to the load R. By feeding the PWM control pulse to the switch tube, the switch state of the switch tube changes continuously until the output voltage in the circuit is higher than the input voltage. In the continuous current working mode, the stable working point of the boost circuit satisfies the inductive volt-second balance principle:

[0095] V S DT=(U C -V S )(1-D)T

[0096] Where: V S and U C are the input and output voltages of the boost converter, respectively, D is the duty cycle of the switching signal, and T is the period of the switching signal;

[0097] For a boost converter, the input power and output power of the converter are equal under steady-state operating conditions, and the relationship between the two can be expressed as follows:

[0098]

[0099] Where: i s is the input current of the boost converter, i o is the output current of the boost converter and R is the load resistance.

[0100] Combined with the relationship between the input and output voltage of the boost converter:

[0101]

[0102] In the steady-state operation of the boost circuit, when the switch is turned on, the output capacitor C is responsible for supplying power to the load during the charging phase of the energy storage inductor. The relationship between the energy released by the energy storage inductor and the corresponding change in the output voltage of the energy storage inductor during the switch-on period can be expressed as:

[0103]

[0104] Where: ΔU C is the ripple of the output voltage of the boost converter, and U is the inductor voltage;

[0105] S8. In the continuous conduction switch mode, there is a negative feedback control loop for voltage and current. The error value is obtained by comparing the output voltage with the reference voltage, and the reference input of the current loop is obtained by calculation by the BP-PI controller. By letting the input current track the reference input, the current signal and voltage sampling phase on the input side are made. The PI algorithm used has the following formula:

[0106]

[0107] e(t)=U k -I ref

[0108] Where: K P Indicates the ratio, K i They represent the integral coefficient, e(t) is the input error, U k is the output of the BP-PI controller, I ref is the feedback value of the current loop, u(t) is the output of the current controller, r t , given input voltage value; e t , the difference between the given voltage and the actual voltage; e t-1 , the difference between the given voltage and the actual voltage at the last sampling moment; y t , feedback value of actual voltage; y t-1 , the feedback value of the actual voltage at the last sampling moment; U t , the output of the outer loop BP-PI controller; U t-1 , the output of the outer loop BP-PI controller at the previous sampling time; K P , K i and U t There are three output layer neurons.

[0109] The parameter value of PI cannot be negative. The excitation function of the network output layer uses a non-negative Sigmoid function, and its formula is as follows:

[0110]

[0111] Where: g(x) is the hyperbolic tangent function.

[0112] The excitation function of the hidden layer neurons is allowed to be negative, and the Sigmoid function is selected to be positive and negative symmetric. The formula is as follows:

[0113]

[0114] Take the performance indicator function as an example:

[0115]

[0116] Where: E(k) is the performance indicator function (loss function), which represents the error between the network prediction value and the target value, r t is the target value, y t is the output (prediction value) of the network.

[0117] The input layer of the network can be shown as follows:

[0118]

[0119]

[0120] in: is the first layer input of the neural network, M is the number of input layers, and in the present invention, M is 8. is the bias node (constant term) of the input layer.

[0121] The input and output of the network hidden layer can be shown as follows:

[0122]

[0123]

[0124]

[0125] in: is the input of the i-th node in the hidden layer, is the weight between the input layer and the hidden layer, is the output of the input layer, is the output of the i-th node in the hidden layer, q is the number of nodes in the hidden layer, is the bias node of the hidden layer (constant term).

[0126] The input and output of the network output layer are as follows:

[0127]

[0128]

[0129]

[0130]

[0131]

[0132] in: is the input of the lth node in the output layer, is the weight between the hidden layer and the output layer, (k) is the output of the hidden layer, is the output of the lth node in the output layer, is the output variable of the network, K P is the control parameter, K t is the temperature related parameter, u k For the control signal.

[0133] Based on the above analysis, the learning formula for the network output layer weight is:

[0134]

[0135]

[0136] in: is the weight adjustment between the lth node of the output layer and the lth node of the hidden layer, α is the momentum term, η is the learning rate, is the error signal of the output layer, e(k) is the error of the loss function, and sgn is the sign function used to determine the gradient direction.

[0137] Learning algorithm for hidden layer weight coefficients:

[0138]

[0139]

[0140] in: It is the error signal of the hidden layer, which represents the combination of the output error of the hidden layer node and the corresponding weight;

[0141] S9. A simulation model is established for the PI double closed-loop control based on the BP neural network to simulate the boost circuit. The BP neural network algorithm is added to the outer loop of the simulation voltage. The inner current loop uses the basic PI module provided by MATLAB. The inner loop controller outputs a periodic duty cycle signal to control the IGBT.

[0142] S10. Finally, the simulation results are analyzed. The simulation target voltage is 30V, the cycle is 0.05s, the output voltage is U0, and the efficiency is calculated by the following formula:

[0143]

[0144] Where: output power is P out , the input power is P in , the efficiency is η.

[0145] Comparison of simulation results within a cycle shows that the efficiency results of Buck circuit simulation show that the peak efficiency reaches 2.8% at 0.005s, and then shows a downward trend. The charging efficiency of the Boost circuit using BP-PI dual closed-loop control is significantly better than that of the Buck circuit. Its efficiency can reach more than 60% within the cycle and has been on an upward trend. It can be concluded that the parameters of the BP-PI controller are adjusted through neural network self-learning to overcome the influence of various adverse factors on the system during operation, thereby better controlling the voltage and improving the charging efficiency.

[0146] The above is only an embodiment of the present invention, and the common knowledge such as the known specific structure and characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A main circuit design based on Boost circuit and a double closed-loop PI control method, characterized in that: The following steps are involved: S1. The receiving end of the wireless charging system receives fixed frequency AC power from the transmitting end of the power grid and converts it into DC power through a rectifier circuit. S2, DC power passes through the DC bus and enters the high-frequency inverter, which changes the current direction by controlling the on and off of S1, S2, S3 and S4, and converts DC power into AC power; S3, the AC power is then modulated by PWM pulse width modulation to change the duty cycle to change the frequency of the AC power, converting the DC power into high-frequency AC power; S4, high-frequency alternating current passes through the primary compensation circuit, and the LCC resonant circuit selects the signal at a specific frequency, amplifies or transmits it, and attenuates the signals of other frequencies; S5, high-frequency alternating current is transmitted to the receiving winding, and the receiving winding then transmits the high-frequency alternating current to the secondary compensation circuit for filtering; S6, the alternating current is then converted into direct current through a rectifier circuit; S7. The DC power is then converted into the required voltage value by the DC / DC converter. The Boost circuit is used. When the switch tube is turned on, the input voltage charges the inductor L and the capacitor C discharges the load R. As time goes by, the current on the inductor increases and the voltage across the capacitor C decreases. When the switch tube is cut off, the inductor L discharges. As time goes by, the current on the inductor decreases and the voltage across the capacitor C increases. At this time, the input voltage and the voltage on the inductor L are superimposed to charge the capacitor C and supply power to the load R. By feeding the PWM control pulse to the switch tube, the switch state of the switch tube is continuously changed until the output voltage in the circuit is higher than the input voltage. S8. In the continuous conduction switch tube operation mode, there is a negative feedback control loop for voltage and current. The error value is obtained by comparing the output voltage with the reference voltage, and the reference input of the current loop is obtained by calculation by the BP-PI controller. By making the input current track the reference input, the current signal and voltage sampling phase on the input side are adjusted. S9, establishing a simulation model for PI double closed-loop control based on BP neural network, and performing boost circuit simulation; S10. Finally, analyze the simulation results.

2. A main circuit design and dual closed-loop PI control method based on a Boost circuit according to claim 1, characterized in that: In S1, the rectifier circuit adopts a bridge rectifier circuit of an ultrafast recovery diode. When the coupling factor between the transmitter and the receiver is high, the signal can be approximated to contain only the fundamental frequency, and the full-wave signal becomes a half-wave after passing through the rectifier bridge.

3. The main circuit design and double closed-loop PI control method based on Boost circuit according to claim 1 is characterized in that: In S4, the bilateral LCC resonant circuit is composed of an LCC resonant compensation circuit composed of a capacitor and an inductor and a magnetic coupling resonant coil.

4. The main circuit design and double closed-loop PI control method based on Boost circuit according to claim 1 is characterized in that: In S7, in the continuous current working mode, the stable working point of the boost circuit satisfies the inductive volt-second balance principle: V S DT=(U C -V S )(1-D)T Where: V S and U C are the input and output voltages of the boost converter, respectively, D is the duty cycle of the switching signal, and T is the period of the switching signal; For a boost converter, the input power and output power of the converter are equal under steady-state operating conditions, and the relationship between the two can be expressed as follows: Where: i s is the input current of the boost converter, i o is the output current of the boost converter, R is the load resistance; Combined with the relationship between the input and output voltage of the boost converter: In the steady-state operation of the boost circuit, when the switch is turned on, the output capacitor C is responsible for supplying power to the load during the charging phase of the energy storage inductor. The relationship between the energy released by the energy storage inductor and the corresponding change in the output voltage of the energy storage inductor during the switch-on period can be expressed as: Where: ΔU C is the ripple of the output voltage of the boost converter, and U is the inductor voltage.

5. The main circuit design and double closed-loop PI control method based on Boost circuit according to claim 1 is characterized in that: In S8, the PI algorithm used has the following formula: e(t)=U k -I ref Where: K P Indicates the ratio, K i They represent the integral coefficient, e(t) is the input error, U k is the output of the BP-PI controller, I ref is the feedback value of the current loop, and u(t) is the output of the current controller; The parameter value of PI cannot be negative. The excitation function of the network output layer uses a non-negative Sigmoid function, and its formula is as follows: Where: g(x) is the hyperbolic tangent function; The excitation function of the hidden layer neurons is allowed to be negative, and the Sigmoid function is selected to be positive and negative symmetric. The formula is as follows: Take the performance indicator function as an example: Where: E(k) is the performance indicator function (loss function), which represents the error between the network prediction value and the target value, r t is the target value, y t is the output of the network (prediction value); The input layer of the network can be shown as follows: in: is the first layer input of the neural network, M is the number of input layers, and in the present invention, M is 8. is the bias node of the input layer (constant term); The input and output of the network hidden layer can be shown as follows: in: is the input of the i-th node in the hidden layer, is the weight between the input layer and the hidden layer, is the output of the input layer, is the output of the i-th node in the hidden layer, q is the number of nodes in the hidden layer, is the bias node of the hidden layer (constant term); The input and output of the network output layer are as follows: in: is the input of the lth node in the output layer, is the weight between the hidden layer and the output layer, (k) is the output of the hidden layer, is the output of the lth node in the output layer, is the output variable of the network, K P is the control parameter, K t is the temperature related parameter, u k is the control signal; Based on the above analysis, the learning formula for the network output layer weight is: in: is the weight adjustment between the lth node of the output layer and the lth node of the hidden layer, α is the momentum term, η is the learning rate, is the error signal of the output layer, e(k) is the error of the loss function, and sgn is the sign function used to determine the gradient direction; Learning algorithm for hidden layer weight coefficients: in: It is the error signal of the hidden layer, which represents the combination of the output error of the hidden layer nodes and the corresponding weights.

6. The main circuit design and double closed-loop PI control method based on Boost circuit according to claim 1 is characterized in that: In S9, the BP neural network algorithm is added to the outer loop of the simulation voltage, the inner current loop uses the basic PI module provided by MATLAB, and the inner loop controller outputs a periodic duty cycle signal to control the IGBT.

7. The main circuit design and double closed-loop PI control method based on Boost circuit according to claim 1 is characterized in that: In S10, the simulation target voltage is 30V, the cycle is 0.05s, the output voltage is U0, and the efficiency is calculated by the following formula: Where: output power is P out , the input power is P in , the efficiency is η.

Images